Educational game designs must balance the often conflicting values of game designers, instructional designers, and content experts. In order to reach this balance, however, colleagues should adopt development strategies that already inform the design of commercial computer games. Commercial game designers recognize that great games are not created by simply envisioning a game, writing a requirements document, and then building it precisely to match the initial specifications. They further recognize that iterative design informed by frequent playtesting is critical to providing an engaging play experience. In illustrating how such practice can inform a collaborative approach to educational game design, Brian Winn and Carrie Heeter outline the development of an educational game designed to meet national science education standards for 7th, 8th, and 9th grade students. They illustrate how playtesting serves an even broader role in educational game design by helping resolve disagreements between gameplay, pedagogy, and content experts. In the project they describe, each group of experts holds some absolute veto power (e.g., the science cannot be wrong); other areas of disagreement, however, require additional negotiation and testing before they can be resolved. Winn and Heeter outline this process of development through three prototypes of the learning game, and they delineate the nine key decisions that resulted in a final version that most successfully resolved the aims of project participants while receiving the most positive feedback from student testing groups. (Contains 10 figures and 4 exhibits.)